Directional Clustering

Directional clustering of vector fields on meshes.

Warning: Major refactoring in progress! Check the develop branch for the latest updates.

Introduction

The initial motivation of this work revolved around principal stress fields. In principle, they suggest good directions to orient material efficiently in architectural structures. This implies that by following these directions, less material would be used to achieve a target level of structural performance.

Principal stress fields are ubiquitously computed by off-the-shelf FEA software and are represented as a cloud of vectors (i.e. a vector field).

As principal stress fields are heterogeneous and form continuous curvilinear trajectories, it is actually difficult for fabrication reasons to place material (in the form reinforcement bars or beams) in a way that exactly match the field directions. It is almost cumbersome, and this is probably one of the reasons why we actually keep on building with orthogonal grids everywhere (take a look at the room around you, for example).

In this work we question the heterogeneity of a principal stress field and inquiry on how much we can simplify it so that we can maximize fabricability while compromising as little as possible in structural performance. In short, what we want is to find the lowest possible amount of different vectors that encode the maximum amount of directional information about a principal stress field. We leverage clustering methods to this end.

Zooming out a bit, this work can be extended to consider other non-structural vector fields such as curvature, or combinations thereof.

This work was initiated by Rafael Pastrana in the School of Architecture at Princeton University in 2020.

Getting Started

Check out the introductory example in the scripts folder.

Installation

The best way to install directional_clustering is to build it from source after cloning this repo.

First, we would need to install the latest version of Anaconda. Anaconda will take care, among many other things, of installing scientific computing packages like numpy and matplotlib for us.

Next, let's create a new conda environment from your command line interface (your terminal on macOS or from the anaconda prompt on windows). The only required dependencies are python, numpy, sklearn, and compas.

conda create -n clusters python=3.7 COMPAS=0.16.9 scikit-learn
conda activate clusters

We should clone directional_clustering from this repository. If you are a macOS user and want to put it in your home folder:

cd ~
git clone https://github.com/arpastrana/directional_clustering.git

Next, let's move into the the repository's folder (the one we've just cloned) and install directional_clustering as an editable package from source using pip:

cd directional_clustering
pip install -e .

To double-check that everything is up and running, still in your command line interface, let's type the following and hit enter:

python -c "import directional_clustering"

If no errors occur, smile 🙂! You have a working installation of directional_clustering.

Contributing

Pull requests are welcome!

Make sure to read the contribution guide. Please don't forget to run invoke test in your command line before making a pull request.

Issue tracker

If you find a bug or want to suggest a potential enhancement, please help us tackling it by filing a report.

License

MIT.